The company is a poster child for advances in AI, but there may be a new sheriff in town.
There’s no question it’s Nvidia (NVDA 1.99%) is a major beneficiary of recent advances in artificial intelligence (AI). The company’s graphics processing units (GPUs) quickly became the gold standard for generative AI, capturing a staggering 92% share of the data center GPU market, according to market research firm IoT Analytics. Nvidia leveraged that advantage to deliver five consecutive quarters of triple-digit year-over-year sales and profit growth.
Many competitors have tried to keep up with the company’s relentless pace of innovation, but none have succeeded. Just this year, NVIDIA revised its product release pace from two years to every year, making competition even tougher for rivals.
But recent entrants into the AI market are making waves, and this could be the first real competition Nvidia faces.
Image source: Getty Images.
challenger
Cerebras Systems is an AI company founded in 2016 that has recently heard rumors of an IPO. The company believes that “AI is the most innovative technology of our generation.”
Cerebras has developed the Wafer-Scale Engine (WSE), a semiconductor giant that takes a different approach to accelerating AI. The WSE boasts 4 trillion transistors and integrates 900,000 compute cores and 44 gigabytes of static random access memory (SRAM) on the chip itself.
Cerebras claims that its unique structure reduces latency, or delay due to data transmission, making the 3rd generation WSE “the world’s fastest commercial AI training and inference solution.” In August, Cerebras announced what it called “the world’s fastest AI inference.” The company claims this is 20 times faster than Nvidia’s GPU-based solutions at a fraction of the cost.
Cerebras updated its claims in a press release issued last week, saying it had “tripled our industry-leading inference performance and set a new record.” The company said tests using Llama 3.2, Meta Platforms’ recently upgraded generative AI model, were “16x faster than known GPU solutions and 68x faster than hyperscale clouds.”
Will this spell be a problem for Nvidia?
While there is some overlap between Nvidia and Cerebras’ AI-focused efforts, it’s important to take a step back and put the competition in context.
Nvidia’s chips have a track record dating back 25 years and have stood the test of time. These GPUs dominate a variety of tasks and markets, including video game graphics cards, data centers, the nascent field of AI, and more recently generative AI.
Beyond the processor itself, Nvidia has taken a more holistic approach, creating software, switches, links, and even an entire plug-and-play system that work together to accelerate processor performance. Additionally, while Nvidia has deep roots in the corporate world, Cerebras is a relative newcomer these days. Relatively easy to deploy, Nvidia’s AI solutions are easy for businesses to implement.
This poses a challenge for Cerebras, as potential customers will have to redesign their systems to incorporate the company’s technology. The associated switching costs could be significant, which could act as a competitive moat for Nvidia. Additionally, companies are less willing to spend large amounts of money on technology that is unproven and has not yet stood the test of time.
Finally, there is the issue of customer breadth. Although Nvidia counts many of the world’s most famous companies as customers, an estimated 46% of its revenue comes from just four customers. Nvidia has been silent on who they are, but it is widely believed to be Alphabet, Amazon, Meta Platforms, and Microsoft.
Meanwhile, Cerebras generated 83% of its 2023 revenue from just one customer (G42 in the United Arab Emirates), which accounted for 87% of its revenue in the first half of this year. A change in direction or a falling out between the two companies could put Cerebra in dire straits and put its other customers (a small number) in a difficult position.
Perhaps more worrying is the fact that U.S. lawmakers have expressed concerns about G42, citing the company’s “extensive business relationships with Chinese military companies, state-owned entities, and Chinese intelligence agencies.” This history and concerns from U.S. regulators could limit Cerebra’s ability to do business with G42 and harm its future prospects.
forest for trees
To be clear, Cerebras is noteworthy because it offers a unique solution that brings Nvidia to a new level of competition that other rivals have yet to achieve. However, before it becomes a serious challenge for Nvidia, the company needs to clear quite a few hurdles.
Cerebras makes a number of claims that still need to be verified. Customer demand will ultimately determine whether Cerebras has what it takes to take on Nvidia.
But until then, Nvidia remains the king of the AI revolution. Although the company’s sales are now about 34 times higher than next year’s, Nvidia’s long track record of success, industry dominance, and strong position have earned it a reputation as a winning company.
John Mackey, former CEO of Amazon subsidiary Whole Foods Market, is a member of the Motley Fool’s board of directors. Alphabet executive Suzanne Frye is a member of The Motley Fool’s board of directors. Randi Zuckerberg is a former head of market development and spokesperson at Facebook, sister of Meta Platforms CEO Mark Zuckerberg, and a member of the Motley Fool’s board of directors. Danny Vena has held positions at Alphabet, Amazon, Meta Platforms, Microsoft, and Nvidia. The Motley Fool has positions in and recommends Alphabet, Amazon, Meta Platforms, Microsoft, and Nvidia. The Motley Fool recommends the following options: A long January 2026 $395 call on Microsoft and a short January 2026 $405 call on Microsoft. The Motley Fool has a disclosure policy.
